Biochemistry 1985, 24, 949-953
949
Reaction of Triosephosphate Isomerase with L-Glyceraldehyde 3-Phosphate and Triose 1,2-Enediol 3-Phosphatet John P. Richard* Institute of Cancer Research, Fox Chase Cancer Center, Philadelphia, Pennsylvania I91 I I Received June 4, I984
Triosephosphate isomerase catalyzes the isomerization and/or racemization reactions of Lglyceraldehyde 3-phosphate (LGAP), the enantiomer of the physiological substrate. The reaction is inhibited by the active site directed reagent glycidol phosphate. The amount of protonation product formation catalyzed by a fixed enzyme concentration is nearly independent of increasing steady-state concentrations of triose 1,Zenediol3-phosphate caused by buffer catalysis of LGAP deprotonation. Therefore, enzymatic protonation of the enediol or enediolate, which could account for the observed enzymatic catalysis of LGAP isomerization and/or racemization, is a t best a minor reaction. Instead LGAP reacts directly a t the enzyme active site. Triosephosphate isomerase catalysis of the protonation of triose 1,2-enediol3-phosphate was expected because of the strong evidence supporting an enediol reaction intermediate for the overall reaction catalyzed by isomerase. The most reasonable explanation for the failure to observe enzymatic protonation is that in solution the enediol undergoes p elimination of phosphate ( t l l Zis estimated to be 10” s) faster than it can diffuse to and form a complex with isomerase. ABSTRACT:
Considerable experimental evidence has been accumulated supporting an enediol(ate) intermediate for the aldose-ketose isomerization reaction catalyzed by triosephosphateisomerase. (1) The enzyme catalyzes the rapid solvent exchange reaction of a proton a to the carbonyl group of dihydroxyacetone phosphate (DHAP) or D-glyceraldehyde 3-phosphate (DGAP) (Rider & Rose, 1959). This result rules out a direct hydride transfer mechanism. (2) Triosephosphate isomerase catalyzes both the isomerization and elimination reactions of triose phosphates (Iyengar & Rose, 1981). This is consistent with an enediolate intermediate, because in solution the enediolate partitions between protonation and p elimination of phosphate (eq 1) (Richard,
cis and t r a n s isomers: 0-1 and 0-2 oxyanions
Ika
ONc”
I
t Po-:
1984; Hall & Knowles, 1975; Bonsignore et al., 1973). (3) The nonenzymatic isomerization reaction in water is through an enediolate intermediate (Richard, 1984). The same mechanism will be favored for the enzymatic reaction because it requires a smaller enzymic rate acceleration than other mechanisms. In order to study the reaction of the enediol(ate) intermediate with isomerase, this species has been generated nonen+This work was supported by US. Public Health Service Grants GM-20940,CA-06927, and RR-05539and also supported by an appropriation from the Commonwealth of Pennsylvania. *Address correspondence to this author at the University Chemical Laboratories, Cambridge University, Cambridge, England CB2 1EW.
0006-2960/85/0424-0949$01.50/0
zymatically by deprotonation of L-glyceraldehyde 3-phosphate (LGAP) (the enantiomer of the enzymatic substrate). Triosephosphate isomerase catalysis of the second half of the isomerization reaction, Le., enediol(ate) protonation, will divert the reaction of the intermediate from the elimination pathway to the isomerization and/or racemization pathways (eq 1). The net increase in the fraction of the intermediate partitioning to protonation products will be manifested experimentally as enzymatic catalysis of the isomerization and/or racemization reactions of LGAP. The results reported here show that triosephosphate isomerase catalyzes the isomerization and/or racemization reactions of LGAP. However, it is also shown that the mechanism for this reaction requires that LGAP react directly with the enzyme in a manner comparable to the reaction of DGAP. Enzymatic protonation of the enediol(ate) generated from LGAP is at best a minor reaction. MATERIALS AND METHODS Most of the materials and methods used here have been reported previously (Richard, 1984). Rabbit muscle triosephosphate isomerase (9000 units/mg) was from CalbiochemBehring. The enzyme was shown to be nearly homogeneous by sodium dodecyl sulfate (SDS) gel electrophoresis but contained a trace impurity with a molecular weight of 15 000. The enzyme subunit concentration was calculated from the absorbance at 280 nm and an extinction coefficient of 3.1 X lo4 M-’ cm-’ (McVittie et al., 1972). Triosephosphate isomerase was assayed by coupling DGAP isomerization to NADH oxidation with a-glycerol-phosphate dehydrogenase. Triosephosphate isomerase was inactivated with DL-glycidol phosphate (Rose & O’Connell, 1969; Schray et al., 1973) by the following procedure. Two milligrams of the barium salt of DL-glycidol phosphate was converted to the free acid by treatment with Dowex 50 and then incubated at 37 OC with 1.5 mg of triosephosphate isomerase in 0.3 mL of 0.1 M triethanolamine hydrochloride (pH 7.9) for 24 h. At this time the enzyme retained 0.3% of its original activity. The isomerization and racemization reactions of LGAP in the presence of triosephosphate isomerase were followed by
-
0 1985 American Chemical Society
950
BIOCHEMISTRY
RICHARD
coupling the formation of DHAP to NADH oxidation with a-glycerol-phosphate dehydrogenase (Richard, 1984). Reaction mixtures of increasing triosephosphate isomerase concentration were prepared by replacing a stock 0.1 M triethanolamine hydrochloride buffer solution with triosephosphate isomerase which had been dialyzed against the same stock solution. Reactions (1.0 mL final volume) were initiated by the addition of LGAP (as a 0.2 M solution at pH 7), NADH (enough of a 0.1 M solution to give a final absorbance of -0.6 at 340 nm), and 10 pL ( - 5 mg/mL) of a-glycer01-phosphate dehydrogenase. The initial velocity of NADH oxidation was determined for the first 2-10 min of reaction, and kis,,+rac values were calculated from eq 2 ([LGAP], is the initial concentration of LGAP). It was shown in control
D H A P t DGAP
methylglyoxal
t Pi
Scheme I1
+
DHAP
DGAP
DHAP
t DGAP
methylglyoxal
t
Pi
d[NADH]/dt
kiso+rac =
[LGAPIo
(2)
experiments that there was no oxidation of NADH in the absence of a-glycerol-phosphate dehydrogenase and that k+mc was constant for a 2-fold increasse in a-glycerol-phosphate dehydrogenase concentration. Values of kiso+rac were reproducible to f 10%. The rate constants for the uncatalyzed and the buffercatalyzed elimination reactions of [32P]LGAP were measured as described previously (Richard, 1984). The reaction in the presence of triosephosphate isomerase was in a volume of 0.1 mL and contained 2 p L of a-glycerol-phosphate dehydrogenase, 2 mM [32P]LGAP(5 X lo7cpm/pmol), and 2 mM NADH. Aliquots of 4 pL were withdrawn and monitored for decreasing organic phosphate (Richard, 1984). RESULTS The presence of 10 mM LGAP had no effect on the observed rate of DHAP formation by the triosephosphate isomerase catalyzed reaction of 1 mM DGAP [the DGAP K, is 0.3 mM (Reynolds et al., 1971)]. However, when the observed rezction velocity was corrected for DHAP formation by the nonenzymatic isomerization of LGAP (Richard, 1984), it was found that isomerase was inhibited 10% by 10 mM LGAP. Figure 1 shows the dependence of the observed rate constants for the formation of DHAP and/or DGAP from LGAP, k,,+,,, on the concentration of triosephosphate isomerase for reaction at 37 OC in 0.1 M triethanolamine hydrochloride (pH 7.9) and 8 mM LGAP. The slope of this plot, kobsd,is 0.23 M-' s-l. Enzymatic catalysis occurs at the active site, because no activity is observed for glycidol phosphate inactivated enzyme (open circle in Figure 1). Previous work in this laboratory was at ionic strength of 1.0 (KC1) (Richard, 1984). Large KCl concentrations were not used here because triosephosphate isomerase is inhibited 10-fold by 0.9 M KCl. Data for enzymatic catalysis of LGAP reaction at 0.0,0.04, and 0.08 M quinuclidinone buffer (pH 7.8) and 2 mM LGAP are given in Figure 2. In a control experiment it was shown that the activity of triosephosphateisomerase remains constant (15% decrease) as the quinuclidinone buffer concentration is increased from 0 to 0.08 M. The rate constants kelimfor the elimination reaction of LGAP to give methylglyoxal and phosphate (see eq 1 ) under the conditions used in Figure 2 are respectively the following at 0.00,0.04, and 0.08 M quinuclidinone buffer: 1.15 X lo4 s-l, 1.82 X s?, and 2.55 X lo4 s-l. It was previously reported (Iyengar & Rose, 1983) that 2 MMtriosephosphate isomerase causes a 60% reduction in the amount of inorganic phosphate formed from LGAP. This experiment was repeated, and it was found that when the
D H A P t DGAP
methylglyoxal
+
Pi
Scheme IV ko t k ~ 1 8 1
LGAP .=------L
methylglyoxal
+
PI
enediolate
enediol
-
DHAP
+
DGAP
elimination reaction of LGAP (2 mM) is carried out at pH 7.9 (0.1 M triethanolamine hydrochloride) and 37 OC in the presence of a-glycerol-phosphate dehydrogenase and NADH to trap isomerization products, the amount of Pi formation decreased only 10% and the rate constant kelimfor the LGAP elimination reaction remained constant [ 1.1 X lo4 s-l (f5%)] as the isomerase concentration was increased from 0 to 20 pM.l There is good agreement between the increase in isomerization and/or racemization products (6%, calculated from Figure 2) and the decrease in elimination products (10%) observed as the isomerase concentration is increased from 0 to 20 pM. Dr. Irwin Rose has also reexamined his previously published work and has obtained results in agreement with those reported here (Rose, 1984a). DISCUSSION Mechanism for Enzymatic Catalysis of LGAP Reaction. Figures 1 and 2 show that triosephosphate isomerase catalyzes the formation of DHAP and/or DGAP from LGAP. Catalysis is only observed at high isomerase concentrations: M isomerase corresponds to 2000 units/mL for enyzmatic catalysis of DGAP isomerization. In previous work it was shown that in solution LGAP undergoes essentially irreversible uncatalyzed deprotonation to give an enediolate which partitions between fast /3 elimination of phosphate and 1100 times slower uncatalyzed protonation to give isomerization (DHAP) or racemization (DGAP) products (see eq 1) (Richard, 1984). It was also shown that LGAP deprotonation is strongly general base catalyzed, that enediolate protonation is general acid catalyzed, but that phosphate expulsion from the enediolate I An attempt to measure kaimat 2 5 0 pM isomerase was unsuccessful, because under these conditions the reaction did not give a stable end point (- 10-20-h reaction time). The slow end point drift may have been due to a-glycerol phosphate hydrolysis by a small amount of phosphatase contaminant in the triosephosphate isomerase.
REACTIONS O F LGAP A N D ENEDIOL 3-PHOSPHATE
VOL. 24, NO. 4, 1985
951
Table I: Buffer Dependence of Triosephosphate Isomerase Catalysis of L-Glvceraldehvde 3-Phos~hateReaction'
[TPI]/
Triosephosphate isomerase catalysis of the reaction of 8 mM LGAP at 37 O C and pH 7.9 (0.1 M triethanolamine hydrochloride). The open circle is for glycidol phosphate inactivated isomerase. FIGURE 1:
0.38 1.oo 1.15 1.oo 0.00 0.97 1.58 0.37 0.04 1.82 0.46 1.21 2.55 2.22 0.08 "At pH 7.8, 37 "C. bThe total buffer concentration. CTheobserved rate constant for the elimination reaction of LGAP keIim= ko + kB[B] (Richard, 1984). dThe increase in the slopes of Figure 2 required if the enzymatic mechanism is that given in Schemes I11 or IV (see eq 9). The values are calculated from the increasing kclimvalues (see footnote c ) . CTheslopes of the lines in Figure 2. /The koM at the given base concentration divided by kOw at zero base. servation that kelimremains constant while kiso+racincreases 4-7-fold (Figures 1 and 2) as the isomerase concentration is increased from 0 to 20 pM rules out the mechanism in Scheme
I. The dependence of kiso+racon buffer and enzyme concentration for Schemes 11-IV is given by eq 3-5, respectively.
-
These equations are derived by assuming a steady-state enediolate concentration, with the simplification that k, k, k,, kBH[BH] kE[E]. This simplification is made because at 10.2 M quinuclidinone buffer concentrations enediolate partitioning to elimination products (k,) is very much faster than partitioning to isomerization or racemization products (k,, kBH[BH] kE[E]) (Richard, 1984). The term fc in eq 4 and 5 is the fraction of enediolate present as the geometric isomer which is the substrate for enzymatic reaction (cis or trans) (Rieder & Rose, 1959; Rose, 1962); it corrects for the different concentrations of the reactive intermediate in the nonenzymatic reaction (total enediolate) and the enzymatic reaction (single geometric isomer). The term A in eq 4 contains a collection of rate constants that are independent of enzyme concentration. The velocity of the reaction between isomerase and the intermediate will be first order in isomerase concentration because the enzyme (104-104 M, Figures 1 and 2) is present in great excess over the enediol and enediolate. The steadystate enediol concentration is equal to
+ +
+
+
+
1
2
M
[TPI] / l O - 5 2: Triosephosphate isomerase catalysis of the reaction of 2 mM LGAP at 37 O C and pH 7.8 (triethanolamine) in the presence of 0 (e),0.04 (O), and 0.08 M (0)quinuclidinone buffer. The reaction mixtures were prepared by replacing 0.1 M triethanolamine hydrochloride with 0.1 M quinuclidinone hydrocliloride. The solid lines FIGURE
are for the calculated experimental slopes, and the dashed lines are for theoretical slopes calculated in Table I (see text) for the mechanisms given in Schemes 111 and IV. is not buffer catalyzed (Richard, 1984). These observations are used as the framework for the four possible mechanisms for enzymatic catalysis of LGAP reaction given in Schemes I-IV. In Schemes I and I1 the enzyme reacts directly with LGAP, to catalyze either enediolate formation (Scheme I) or the overall reaction (Scheme 11). In Schemes I11 and IV the enzyme reacts with an intermediate generated by nonenzymatic deprotonation of LGAP, either the enediolate (Scheme 111) or the enediol (Scheme IV). In all these schemes a broken arrow has been drawn for enediolate return to LGAP, because LGAP deprotonation is essentially irreversible (Richard, 1984). In Scheme I triosephosphate isomerase acts to increase the rate of enediolate formation. Enzymatic catalysis of enediolate formation must lead to the same proportional increase in the observed rate constants for the appearance of all products formed by partitioning of the enediolate. Therefore, the ob-
+
[ ( k o kB[B])[H]/Kak,][LGAP]
+
-
-
lo-' M
where ko k,[B] I2.55 X lo4 s-l (Table I), [LGAP] I 0.008 M, [HI = lo-'.' M, K, M (Chiang et al., 1984), and ke 5 X lo5 s-' (see the following section). The steady-state enediolate concentration will be smaller still because the enediol [pKa 10.9 (Chiang et al., 1984)] is largely protonated at pH 7.9. The dependence of the observed second-order rate constants for isomerase reaction, kow = d(kiso+rac)/d[E],on buffer general base concentration are given for Schemes 11-IV by eq 6-8, respectively. The normalized dependence of k&sd on general base Concentration for Schemes 111 and Iv (k&d at a given base concentration divided by kobd at zero base) is given by eq 9.
-
-
952
RICHARD
BIOCHEMISTRY
The ratio kobsd/(kobsd)O in eq 9 is also equal to the ratio of the observed LGAP elimination rate constants (kelim)at the two buffer concentrations, since kelimis essentially equal to the rate constant for substrate deprotonation: ko kB[B] (Richard, 1984). Schemes 111 (eq 7) and IV (eq 8) require a first-order dependence of kokdon base concentration because the effect of general base catalysis is to increase the steady-state concentration of the intermediate which reacts with isomerase, either the enediolate (Scheme 111) or the enediol (Scheme IV). Scheme I1 (eq 6) predicts a zero-order dependence of koM on [B], because here the reactions of the general base and enzyme are with different species along separate pathways. Data for the reaction of LGAP at increasing quinuclidinone buffer concentrations are summarized in Table I. The observed second-order rate constants for the enzyme-catalyzed reaction, koM,are nearly independent of buffer concentration, whereas the largest required increase in koM for Schemes I11 and IV,calculated from the increase in kellm(see above), is 120%. The dashed lines in Figure 2 show the increase in koM required for Scheme I11 or IV. The 20% increase in kobsdat 0.08 M quinuclidinone is just within the estimated experimental error of f 10%; it is consistent with a >80/ k,[E] for Scheme V. The value for k, may be estimated by comparison of the reactions of the enols of acetone and dihydroxyacetone phosphate. A value of 50 s-I (25 "C) for acetone enol protonation at pH 7.9 is calculated from the rate equation for protonation at pH 10-12 (Chiang et al., 1984). The same value is used for the water protonation of the enediol at pH 7.9 and 37 " C because of the offsetting corrections required for the different reaction temperatures and the slightly greater thermodynamic stability estimated for the C-3 DHAP carbanion [pK, N 18 (Richard, 1984)] compared to the acetone carbanion [pK, = 19.2 (Chiang et al., 1984)l. In both cases the primary pathway for enol reaction
* The /cow value at ionic strength of 1.0 (KCI) [calculated from data in Richard (1984)l is about 10 times smaller than the /cow values reported here due to KCI inhibition of isomerase. The small kdw value was more easily determined from the plot with the shallower slope. At pH 7.9 the mechanism for the uncatalyzed deprotonation of LGAP is through intramolecular substrate deprotonation by the phosphate dianion (Richard, 1984).
REACTIONS OF LGAP AND ENEDIOL 3-PHOSPHATE
is protonation of the small pool of the highly reactive enolate by water. The rate constant for enediol phosphate protonation is faster because of intramolecular protonation of the enediolate by the phosphate monoanion. Intramolecular DHAP deprotonation at pH 7.9 is 100 times faster than hydroxide deprotonation (Richard, 1984), and there must also be a 100-fold acceleration of the protonation reaction, because the added intramolecular pathway cannot change the equilibrium constant for substrate deprotonation. The k, value is increased further because the rate of enediol breakdown to inorganic phosphate is 100 times4 faster than the rate of the intramolecular protonation reaction (Richard, 1984). This gives an estimated k, value of 50 s-l X lo2 X lo2 = 5 X lo5 s-l. The value for kE will be equal to or less than the maximum rate constant for diffusional binding to the enzyme, lo9 M-' s-l (Hammes & Schimmel, 1970). Combination of kEwith the largest enzyme concentration used in this study gives a maximum value of k E [ E ]5 (lo9 M-' s-l) (1.5 X lo4 M) I 1.5 X lo5 s-l for enzymatic protonation of the enediol. This is smaller than the estimated elimination rate constant k,. Therefore, it is predicted that 1.5 X 1O-" M isomerase subunits will not trap a large fraction of the enediol, because the lifetime for the enediol in water is not long enough to allow for diffusional encounter with isomerase. It is well recognized that triosephosphate isomerase must act to decrease the rate of enediol breakdown by /3 elimination of phosphate, if the species is an intermediate of the enzymatic reaction (Iyengar & Rose, 1983; Alber et al., 1981). The facile breakdown by /3 elimination of phosphate will make the direct study of the chemical and enzymological reactions of the enediol of DHAP and GAP much more difficult than the study of the reactions of the enol of pyruvate. The latter compound, which is also an important enzymatic intermediate (Rose, 1982), has a half-time for deuteration in D 2 0 a t pD 6.4 and 20 OC of 3.6 min (Kuo et al., 1979). The results reported here are inconsistent with the previously reported enediol phosphate lifetime of s and the reported 50% trapping of the enediol phosphate by 6 X M isomerase subunits (Iyengar & Rose, 1981). The experiments from which these values were calculated have been found to be irreproducible (Rose, 1984b). Direct Reaction of LGAP. A two step mechanism for direct enzymatic catalysis of the isomerization and/or racemization reactions of LGAP is written in Scheme VI. Scheme VI E
+ LGAP
EsLGAP kobsd
5E + DGAP + DHAP
kcat = [LGAP]
+ Kd
Equation 11 gives the relationship between the observed second-order rate constant kobd calculated from the slope of the plots of Figures 1 and 2 (no quinuclidinone) and the kinetic parameters in Scheme VI. The kobsdvalues at 2 (0.38 M-l s-l) and 8 mM (0.23 M-' s-l) LGAP differ by less than a factor of 2, showing that the enzyme is below saturation at 2 mM LGAP. Approximate values of Kd = 7 mM and k,,, = 3.5 X s-l can be calculated from a double-reciprocal form of eq 1 1 and the kobd values at 2 and 8 mM LGAP.' A There is a small KCI salt effect that leads to a 40% increase in the rate constants for LGAP elimination and isomerization as the KCI concentration is decreased from 0.9 to 0 M. The ratio for enediolate partitioning between elimination and isomerization product formation calculated from the ratio of these rate constants is therefore independent of KCI concentration.
VOL. 24, NO. 4, 1985
953
comparison with the steady-state kinetic parameters for reaction of DGAP (Knowles & Albery, 1977) shows an approximate decrease of lo6 and lo9, respectively, in the firstorder rate constant k,, and the second-order rate constant kcat/kMfor the reaction of the nonphysiological substrate LGAP compared to DGAP. Estimated kinetic parameters for 3-quinuclidinone catalysis of LGAP isomerization are Kd = 230 M and k,,, = 0.4 s-l (Richard, 1984). Comparison with the enzymatic values shows that k,,, for the enzymatic reaction is 100 times slower than quinuclidinone deprotonation of substrate within a weak encounter complex, suggesting that catalysis of the enzymatic reaction is due to slow, adventitious substrate deprotonation by a base at the catalytic site. Discussion of the stereochemistry of the enediol(ate) intermediate for enzymatic catalysis of LGAP reaction awaits the determination of the stereochemistry of proton incorporation into C-3 of DHAP. ACKNOWLEDGMENTS I thank Dr. Irwin Rose for his support. Registry No. LGAP, 20283-52-7;triosephosphate isomerase, 9023-78-3;triose 1,2-enediol3-phosphate, 57-04-5. REFERENCES Alber, T., Banner, D. W., Bloomer, A. C., Petsko, G. A., Phillips, D., Rivers, P. S., & Wilson, I. A. (1981) Philos. Trans. R . SOC.London, Ser. B 293, 159. Bonsignore, A,, Leoncini, G., Siri, A,, & Ricci, D. (1973) Ztal. J. Biochem. 22, 131. Chiang, Y., Kresge, A. J., Tang, Y. S., & Wirz, J. (1984) J. Am. Chem. SOC.106,460. Davies, C. S. (1962) Zon Association, pp 77-87, Butterworth, London. Hall, A., & Knowles, J. R. (1975) Biochemistry 14, 4348. Hammes, G. G., & Schimmel, P. R. (1970) Enzymes, 3rd Ed. 2, 67. Hine, J. (1975) Structural Effects on Equilibria in Organic Chemistry, p 120, Wiley, New York. Iyengar, R., & Rose, I. A. (1981) Biochemistry 20, 1229. Iyengar, R., & Rose, 1. A. (1983) J. A m . Chem. SOC.105, 3301. Knowles, J. R., & Albery, W. J. (1977) Acc. Chem. Res. 10, 105. Kuo, D. J., O'Connell, E. L., & Rose, I. A. (1979) J. A m . Chem. SOC.101, 5025. McVittie, J. D., Esnouf, M. P., & Peacocke, A. R. (1972) Eur. J . Biochem. 29, 67. Reynolds, S. J., Yates, D. W., & Pogson, C. I. (1971) Biochem. J. 122, 285. Richard, J. P. (1984) J. A m . Chem. SOC.106, 4926. Rider, S. V., & Rose, I. A. (1959) J. Biol. Chem. 234, 1007. Rose, I. A. (1962) Brookhaven Symp. Biol. 15, 293. Rose, I. A. (1982) Methods Enzymol. 87, 84-97. Rose, I. A. (1984a) J. Am. Chem. SOC.106, 6117. Rose, I. A. (1984b) Biochemistry 23, 5893. Rose, I. A., & O'Connell, E. L. (1969) J. Biol. Chem. 244, 6548. Schray, K. J., O'Connell, E . L., & Rose, I. A. (1973) J. Biol. Chem. 248, 2214. Trentham, D. R., Murray, C. H., & Pogson, C. I. (1969) Biochem. J. 114, 19. The Kd value can be used to calculate that 10 mM LGAP will inhibit by 25% triosephosphate isomerase catalysis of the reaction of 1 mM DGAP [K, for DGAP is 0.3 mM (Reynolds et al., 1971)], whereas only 10% inhibition was observed (see Results). The difference can reasonably be attributed to experimental error in the Kd determination from a two-point double-reciprocal plot.